capitolul 23

Chapter 23

40/20 Meter Transceiver

This is an experiment in creating a transceiver that uses many of thesame parts and can be built for either 40 or 20 meters. Now, like anyexperiment this can work or it can fail. We wont know until the finaltesting is done. But that is what experimentation is all about. You cantlive if you are afraid of dying.

So, let us begin.

23.1 The VFO

The heart of a transceiver is typically a common frequency source. Ifthe source is fixed, typical of a crystal controlled rig, then the transceiverworks on one frequency only. This is the basis premise of the SmallWonder Labs Rock Mite and other similar transceivers.

If the common frequency source can be changed, then we have a vari-able frequency oscillator (VFO) or variable crystal oscillator (VXO) whenthe primary frequency control is a crystal, but circuitry has been addedto vary the crystal frequency. For this transceiver I am going to usethe Colpitts oscillator that Dave Benson, NN1G at the time, used in theNE4040 transceiver that was a kit put out by the New England QRP Clubin 1994. Dave has given me permission to use the design as a teachingaid and for that I thank him.

Here is the schematic for the VFO. Please note. There is a missing con-

277

278 CHAPTER 23. 40/20 METER TRANSCEIVER

nection, noted by the dot and label 9 at the joining of the two wires ofresistor R19 and collector common point for the NPN transistor, 2N39042N2222 2N4401 or others will work here.

23.1. THE VFO 279

Figure 23.1: Colpitts VFO.


Note that I have labeled some of the nodes in the VFO with numbers inblue for ease of discussion.

The VFO is a Colpitts oscillator. I like the colpitts oscillator as it does notrequire a tapped inductor, which for many is a difficult configuration towind on a toroid.

Any resonant frequency of an oscillator is determined by an inductorand capacitor combination that satisfies the formula

=1

2pipLC

(23.1)

Lets start with L. The inductor L in the vfo circuit is a T50-6 toroidwound with 25 turns of #22 wire. If we go to http://www.kitsandparts.comand use a web calculator for toroids we find that the toroid has an in-ductance of 2.50H.

In order to determine the desired VFO frequency, we first need to knowa couple of things. The first is the IF frequency for the receiver andthe second is HF frequency we want to operate at. Lets first startwith the 40 meter band and use the QRP international calling frequencyof 7.030MHz to start our analysis. Because I won a bid on ebay for11.059MHz crystals, I plan on trying to use a lot of them in the futureand that includes this project.

In order to get to 40 meters, I need a frequency that will subtract from11.0592MHz to get to 7.030MHz. That means a frequency of 4.022MHzfor the VFO at that operating point in the HF spectrum.

What capacitance must we see across the L of 2.50H? Plugging intothe frequency formula we get a capacitance value of 626pF.

Now I do not trust myself with a calculator. Too many ways to make amistake and I have to do the calculations at least twice to double checkmy work. Now I just write a quick python program to do the calculation.Here is the code that I used for the above calculation.

l = 2.50e-6f = 4.022e6

c=1.0/(2.0*2.0*3.14159*3.14159*l*f*f)print c

23.1. THE VFO 281

The good thing about a piece of code like the above is that it is handyto come back and modify the values for another frequency. I also havea copy to double check against for programming error. Not likely withsomething this simple. I did take the scientific calculator and did thecalculations and got the same result, so Im not totally incompetentwith a calculator. :-)

23.1.1 Capacitor calculations

We know L, lets see if we can determine C across L in the colpittsoscillator. I am going to neglect resistance in the coil and the wires forthis calculation as we are not going to calculate the Q of the circuit.

Start at node 5 in the schematic. There is C7, L and then C9 and C10(in series) all in parallel. Then from node 5 we pass through throughC6 to three dual capacitor series combinations to ground. This soundscomplicated but here is the schematic for the inductor and capacitorcombination.

Looks complicated, but it really isnt. Just some simple mathematicsand Ill show you the steps here. First. There are two unknowns in theschematic. The first is C7 marked with value XX. This is a cap that youdo not put in until you first test the VFO powered up with all the rest ofthe caps in play. Then, if the frequency of the VFO is running too high,you add a cap to bring the frequency down. Remember the frequencyis inversely proportional to the square root of C and increasing C bringsdown the resulting value.

The varactor diode is shown as D2. If you go to the following web site,you can purchase the cap for about $2 for one including the postage.By buying more you can reduce the cost of a varactor diode down toclose to 85 cents (US currency) per unit. I bought some of the MV1662sthat I have in my parts collection from Earl and he is easy to do businesswith. This was some time ago and he still has a few of the diodes leftas determined by his web page.

http://www.earlandrews.com/MV1662.html


Look at Figure 4.11 for a curve that shows the capacitance vs voltagecurve for this part. Lets pick a value of 100pF for D2 to plug into theformula for the frequency in the LC combination for this VFO.

I am going to do the math from left to right in the figure.

C3 and C2 8.3 pFC8 and D2 45.1 pFC4 and C5 1350.0 pF

Total for the three combos in parallel is 1403 pF.

C6 and combo 984 pF

C9 and C10 9.6pF

Total capacitance is 994 pF.

Figure 23.2: LC network for VFO frequency.

23.1. THE VFO 283

This in parallel with the 2.50H, results in a resonant frequency of3.19MHz, which is higher than expected. For the NE4040 the IF fre-quency was 4.00MHz and thus for 40 meters we would like a value of3.030 MHz to get to the QRP frequency of 7.030MHz. Double checkmy calculations above. Ill leave things as they are and we will buildthe circuit and see how the physical results compare with the theoret-ical calculations. With the C7 cap in parallel, by adding capacitance inparallel we will lower the operating frequency of the VFO and thus canbring it into alignment with the desired frequency range.

Since the VFO frequency with 100 pF for the varactor diode is too highat 3.19MHz, lets go and add one turn to the toroid and make it 26 turnsor an inductance of 2.70H. This brings the resonant frequency downto 3.07MHz, which is closer to the desired value and we can tweek thecircuit after that very easily.

23.1.2 SPICE Simulation of the VFO

Here is the output from a SPICE simulation of the Colpitts oscillator. Iuse ngspice with the linux operating system to do all my simulationwork. I did the simulation as part of the Elmer 101 series that was doneon QRP-L using the SWL-40+ transceiver for a series of tutorials onkit building and understanding of the different building blocks of a rig.What is interesting is that the series of tutorials is online. The analysisof the Colpitts oscillator has an error in determining the C value acrossL1. Hopefully I got it correct here. The graph shows that when the VFOis powered up there is a delay of almost 10 microseconds before theoutput reaches a near stable amplitude.

We do not have to worry about the start up time as the VFO runs con-tinuously when the rig is powered up.

In the VFO section all the caps that we did the calculations with must beeither C0G or NPO so that the VFO does not drift significantly during thewarmup period. When the transceiver is powered up there will be someheat generated do to power losses in resistors and other componentsand this heat will cause some drift and if care is not taken the driftwill be significant and make it difficult to use in day to day operation.Building the VFO and testing it will insure that we do all that we can tomake it stable.


23.1.3 Winding the toroid and measurement

To check the work so far, it is time to build the VFO. First, Ill wind thetoroid and measure its inductance using the AADE L/C Meter II and my

Figure 23.3: Colpitts VFO simulation.

23.1. THE VFO 285

own fixture.

Winding the toroid is fairly easy. Just dont rush the process. The firstthing I see a lot on the Internet is the question about wire size for toroidcoils. The inductance of a coil is determined by the number of turns.The larger the wire size the lower the internal resistance of the lengthof wire used to wind the coil and thus a resulting higher quality factor,Q, for the coil. But, if you do not have the wire size called for, a smallerdiameter sized wire (higher wire size number) will work. A larger diam-eter wire size may not be possible depending upon the number of turnsrequired and the diameter of the toroid.

Note also that the direction in which you wire the toroid, as long asthe number of turns is the same, does not affect the resulting L value.The winding direction does effect the position of the two wires and theirplacement on a PCB. So, if you are building a kit, make sure you lookat the wire placement on the PCB before you wind the toroid to makesure that the wires are aligned correctly when putting into place andsoldering. It is not a happy day when you have to unwind a coil and redoit. There is no hard and fast rule as to which direction toroids should bewound. I prefer what corresponds to what Diz calls counterclockwise.

Here is a URL for a site with a set of nice photographs to illustratethe clockwise and counterclockwise directions done by Diz, W8DIZ, atkitsandparts.com.

http://tinyurl.com/nf2jf2w

23.1.4 Transmitter Mixer

The function of the transmitter mixer to to combine the VFO frequencyand a frequency generated by a crystal oscillator to generate the RFfrequency to be transmitted. This is the reverse of the process of takinga received signal and mixing with the VFO to get the IF frequency.

The process of mixing is going to generate a sum and a difference ofthe two frequencies fed into the mixer and some other nasty products.So, following the mixer we need a bandpass filter to seperate out thedesired transmitter frequency. Here is the schematic of the additionalmixer and bandpass filter.


In the toroid chapter in this document I showed a one wire fixture tosort toroids when -43 and -61 toroid types get mixed up. I am going tobe OCD here and measure some T50-6 forms to match the A values forthe two coils I am going to wind using the two directions. I am going touse #24 wire as I do not have a quantity of #22 wire on hand and I amgoing to do 25 turns (25T) for each coil.

OK. Made a decision here to change the VFO frequency and to go aheadand add the first transmitter mixer in order to kill several birds with onestone. If I am going to prototype the VFO using a muppet board, thenI might as well go ahead a do a board that has another section or twoof the transceiver and build and test in sections. If I use 11.059MHzfor the IF frequency, then work on 20m, then to get to the QRP call-ing frequency of 14.060MHz we need a VFO frequency of 14.060MHz -11.059MHz or 3.001MHz. Using the 994pF capacitance calculation theL1 inductance needs to be 2.83H. Looking at the W8DIZ calculator, wewill need 27 turns on the T50-6 toroid.

23.1. THE VFO 287

Figure 23.4: VFO and transmit mixer followed by bandpass filter.


23.2 First Section Build

OK, now is the moment of truth. I am going to build the transmittersection shown on the previous page. I have gone and added Q2, whichis part of the keying circuit. Ill come back and add the labeling later. Itis not important right now.

The purpose of having the keying circuit turn on and off the transmitmixer is to keep from hearing the local oscillator VFO converted to thefrequency on which we are listening and transmitting on. You and I donot want to hear the signal interfering with any one we are working onthe air.

Figure 23.5: Muppet board for the VFO section.

Now this muppet board is just for the VFO, keying circuit and first trans-mitter mixer. I could have sat down and laid out the entire transceiver,but I have yet to design a driver and PA section for the transmitter. Theoriginal NE4040 and SWL-XX+ series from Dave only put out about 2W.

23.2. FIRST SECTION BUILD 289

In most cases about 1.5W, which for a beginning QRPer is asking toomuch IMHO. You have to have a lot of patience and a good antenna tobe successful. I have done WAS on 40, 30 and 20m with 0.95W and Ican tell you it took a lot of time on the air to get it done. If it was easy,everybody would be doing it.

OK. Here is the photo of the build up to the VFO. I build the voltageregulator section and measured its output at 8.07V. This just to makesure I have everything right. It is of no use to build everything and thenhave to look for problems. I am a big fan of build things one section ata time in logical order and test each section.

Figure 23.6: The VFO section up and running.

I built the VFO section. Powered it up and use a scope to measure theoutput at nodes 6 and 8 in the schematic to make sure the oscillator isworking. I am getting about 25mV out at the input to the mixer of thetransmitter, pin 2 of the NE602A. This is the same value that I get frommeasurements on a working NE4040 transceiver.


The RF voltage level is too low for any frequency counter that I have,so there is no need to adjust the operating frequency of the VFO at thistime is ever for this first experiment. This board is not going to be usedin a final build.

I laid out the board using ExpressPCB and without the use of the schematiccapture routine of ExpressPCB design software. I just take the schematicand use the computer screen to lay down the pads and traces. So, thelayout is not going to be the final version as I learn as I go on just wherethe optimal placement is going to be. And, I dont care if you or anyone else can do better. This is not a competition. It is fun for me andmy aim is to just get a good working rig up and running on the air.

Figure 23.7: The keying section added.

The keying section is fairly straight forward. When the input to thebase of the 2N3906 is pulled to ground (shorted or brought low by akey or keyer) the transistor begins conduction and thus a voltage isapplied to the transmitter mixer. This turns on the NE602 and allowsthe output from the VFO being fed into pin 2 and at the same time

23.2. FIRST SECTION BUILD 291

turns on the local oscillator with the 11.059MHz crystal and feeds thesum and difference of the two frequencies to the output.

The desired frequency is selected by the bandpass filter.

Here is the photo with the mixer installed and working.

Figure 23.8: The transmit mixer section added.

At this point, after looking at the output from the mixer and seeingthat all is working as expected, I have decided not to build the band-pass filter section. It is all passive components and easy to tune to thefrequency range desired. And my supply of variable trimmer caps islimited and I want to not run out of them.


23.3 Drivers and Final Power Amplifier tests

In this section I want to look at a selection of power amplifier circuitsand power transistors to determine what I want to use for both 20 and40 meters.

This will involve some test fixtures and some self destruction tests todetermine what will stand up to some abuse and still survive after-wards.

capitolul 23

Documents